Ligia BARROZO, Universidade de São Paulo, Brazil
João Paulo BARBOSA, Universidade Ibirapuera, Brazil
Nara SILVA, Universidade de São Paulo, Brazil
Alex FLORINDO, Universidade de São Paulo, Brazil
Healthy nutrition and practice of physical activity are known to be protective against noncommunicable diseases (NCD) such as cardiovascular diseases, cancer, obesity, and diabetes. Walking during leisure time or active mobility is an inexpensive way to practice some physical activity. The walkability index (WI) is a powerful measure to help understand how the built environment influences the decision of walking. In theory, the more walkable an area is, more people will be engaged in active mobility. Understanding this relationship can contribute to design innovations in urban planning to help reduce the burden of NCD and improve overall population health. Thus, we calculated the WI in the city of São Paulo, Brazil, adding the relief at the Frank et al., 2010 methodology and compared it with the answers provided by the Origin-Destination Survey 2017, by geographic zone of the city. We applied Multiscale Geographic Weighted Regression using the software MGWR 2.2. Results show that about 27% of the interviewed walk from the origin to their destinations. On average, the time of walking is 12 minutes. The reasons for walking were short distance (89.42%), expensive public transport fares (2.51%), to practice physical activity (2.10%), long distance to the closest public transport stop (0.53%), long wait for public transport (0.47%), slow trip by public transport (0.18%), crowded transportation (0.13%) and others (1.14%). MGWR models using the minimum, mean and maximum WI of the zone present overall associations of Adj. R2 0.41, 0.31 and 0.48, respectively. In the best model, maximum values of WI, there is a local influence (bandwidth of 50 neighbors). There is a significant association in 43.5% of the zones with R2 varying from 0.21 to 0.70 and coefficients varying from -0.97 to 0.84. Although significant and relatively strong in some areas of the city, the WI has a small impact in the decision of walking in face of other reasons included or not in the survey.
Mots clés : Physical activity|built environment|Multiscale Geographic Weighted Regression|Origin-Destination Survey
A105160LB